I had an interesting email today from Parcel2Go.com, who I use occasionally to send bits of cars around the country. It seems that they’re getting into buying links in an interesting way! I wonder what Google makes of this? 
Apologies for the paucity of posts recently, I’ve been very busy delivering training and consultancy for the eBusiness Programme. Presenting my new Google Analytics & Conversion workshop (four times in a week!), Analytics is very much on my mind and I spotted this on Erik Vold’s blog, relating to the previous post here about the Canonical tag and preventing Analytics from double-counting pages:
http://erikvold.com/blog/index.cfm/2009/4/23/relcanonical-and-google-analytics
Erik has created some Javascript and amended ga.js to stop Analytics counting canonical pages in your stats. Very handy! 
The major search engines (Google, Yahoo and MSN of course! ) have announced that they will all support the use of the “canonical” tag as a way to understand duplicate content pages in a site. In essence, this means they are providing a way for webmasters to tell them which version of a page is the “master” and which are “alternative versions”. So, wherever you have the same content appearing in multiple pages on your site (say, a product on an e-commerce site that appears in several categories and hence has multiple addresses), you can tell the search engines that you know these are “duplicates” and which one it should treat as the “master”.
This is done with a simple line in the section <head> of your pages:
<link rel="canonical" href="http://www.website.com/category/product-page">
The canonical tag only works within a domain (including subdomains and folders) and should use absolute URLs rather than relative URLs (that means http://... addresses, not just ../path URLs).
Remember this will only work for pages that are very similar, so don’t think you can go applying the canonical tag to any page just to point some extra link juice somewhere else on your site!
There’s a slightly longer article about using these tags here at Search Engine Land.
I received an amusing email via one of my site’s yesterday whilst feeling slightly feverish on the sofa (it’s December, it must be time for a cold!) The (poorly formatted) email was from a chap offering to “hand deliver… quality links and rankings” to me.
How would one do such a thing, I pondered? He went on to offer a 30 day free trial and guaranteed first page rankings for a particular relevant phrase within 90 days. The free trial bit was what got me twitching – a free trial suggests that the service being provided can be turned off/undone. If we’re talking about links from “high PR [PageRank] quality sites” (as he was), how could these be simply switched off?
The answer is obvious, of course. You own all the sites providing the links. Yep, we’re talking about a link farm – a network of websites created purely and simply to link to others, not to provide useful content to humans. So I went investigating, as this chap handily linked to some of his clients on his website.
What I found was a long list of WordPress-based blogs linking to this client’s site. I have to say, it was well done – every (keyword-rich) link was part of a decent-length blog post about a relevant topic and from my brief surfing, each post on a different blog was unique content. Quite time-consuming to create. What was a big giveaway was the incredibly broad and random spectrum of blog posts on each site, from NLP to office chairs to funeral homes. Not the sort of breadth that an individual’s own blog/site usually covers!
Whether this chap is directly responsible or simply reselling another service, I’m not sure, but as his own site has zero PageRank and a lot of the link farm’s content is about foreign businesses, it’s probably the latter. It wasn’t clear how much this “service” costs.
So why am I irritated by it? Well, firstly, it is a (usually) short-term solution with dire consquences when it falls apart. Either it becomes enormously expensive, because you pay by the month and the moment you do, all those links disappear, or worse, your site gets banned by Google because they are very much against link farms.
It is far better to build genuine links from independent websites, as they will last for a long time and don’t present any danger in terms of Google penalties. One way to find these links is to analyse those of your competitors, a service that I provide. If I found a list of spammy link farm sites in the list of a competitor’s links, I’d be straight to the spam report form on Google. You might call that fighting dirty, but if your competitors are breaking the rules while you strive to meet Google’s ever more stringent criteria, why shouldn’t you level the playing field?
If you have a Google account and are logged in whilst searching, you’re now likely to see these buttons (in the UK at least):
These let you promote (the up arrow) or remove (the cross) an individual search result from the listing. You can also add your own comments using the speech bubble icon at the bottom of each result. The first time you use these you’ll see this message:
So, as you can see, Google has effectively taken a Digg-style approach to its search results, allowing users to vote (and negatively vote) for pages, as well as providing comments on them. Should you remove a search result from the page, it will disappear in a puff of smoke and re-appear at the bottom of the listing, along with some other options:
So, this functionality is useful to users to help them organise their search results, telling Google the type of sites they don’t want and to make notes they can refer to later when searching again for similar information. But will it affect SEO?
I think the answer has to be… possibly. At the moment, I find it unlikely – it’s too early and Google will want to see how users respond and what data is created. If the results are positive, maybe they will incorporate this user feedback into rankings, although it is a very simple system – “yes or no”, in effect, not “how good”. I think the notes are useful, as people can identify spam sites etc. before others click through to them, but whether keywords in these comments will affect rankings is a slightly more difficult proposition.
If Google did let SearchWiki affect rankings, how long before people in, um, less well paid countries than ours started offering “1000 positive reviews from 1000 unique Google accounts” type services? Then Google would have to start checking Google accounts against IP addresses to identify mass activity from a particular location, but even that is quite easy to get around. Ultimately such action might affect the whole user-friendliness of a Google account, which would cause far more harm than good.
No, I think the likely outcome is that SearchWiki will be included as an option for users to get input from others as to whether sites are any good. You may be able to search notes made on SearchWiki directly at some point, essentially mirroring a search on Digg (or other social bookmarking sites). I wouldn’t be surprised if additional social bookmarking features like keyword tagging appear as well; but I don’t think we’ll see this stuff directly affecting rankings for others, as the implications for spamming are too difficult to control.
As a final comment, don’t forget that Google already personalises your search results if you’re logged in – for instance, your own website may well appear higher in searches than other people see it, because you often go there! SearchWiki just takes that to the next level, so rather than Google guessing what you do/don’t like, you can tell it directly.
I’ve been busy training a lot lately and also developing a new keyword analysis tool, which I’m very excited about. It’s nearly ready, I just need to get a few bugs ironed out and come up with a name!
The keyword tool will use both Google and Wordtracker data and perform competitive analysis on a number of factors, making life much easier for those of us trying to find the right target phrases without spending hours in front of the computer.
I was discussing the issues around “hidden” or “protected” content with a client yesterday, specifically the problem that as a website owner you want as much content in the search engine’s index as possible, so that your site will be found, but you don’t actually want humans to see it without registering/paying.
This is an issue that has plagued paid-for content sites for years (see Danny Sullivan’s history lesson here). The problem being that whilst there are pretty simple technical solutions to allowing search engine spiders into your site, whilst preventing access to the casual human browser, pretty much any way of doing this you can come up with constitutes “cloaking” in the eyes of the search engine. If you have a look at Google’s Webmaster Guidelines on the subject, you can understand why this practice is frowned upon – they don’t want users to be taken somewhere they weren’t expecting, as that could severely affect the quality of the user experience and ultimately lead to people using another search engine.
I noticed that Google had made a blog post attempting to deal with this problem while I was on holiday – they want users to be able to find “protected” content because it may be just what they’re looking for, but not at the expense of inviting spam into the index. The solution is simple – allow Googlebot to index your site and when a user finds that page via a Google search, let them see the full page. If they want to access another “protected” page, Google is quite happy for you to require registration/payment; but not for that first page/article they clicked to from the search result. They call it “First Click Free” (FCF), something that has been accepted in Google News search for some time.
Initially, that sounds like a sterling solution. But it doesn’t take long to realise the problems here – firstly, a simple site: command search on Google for the site in question will reveal every page on the site. According to Google’s rules, if you click on any of those pages in the search result, you should see the whole article for free. So, a simple run down the full list of pages provided by that site: search gives you access to every page of paid content on the site in question.
Secondly, there are some simple technologies freely available out there to make you appear to be Googlebot or to make it look like every page you view has been referred from a Google search (here’s just one). So, using these, it would be simple to browse a site conforming to Google’s FCF rules and get access to every page – you wouldn’t even need to keep going back to that site: search listing.
So, what should the webmasters of such sites do? Well, you could take the view that the vast majority of web users have no idea about the site: command, changing user agents or accessing Google’s cache (the “Cache” link that appears under each search result that shows Google’s copy of the page in its database, rather than the “live” page). In which case, the vast majority of your site’s visitors will experience the site just as Google suggests.
However, if this becomes a popular method of allowing Google access to hidden content, how long before tools are developed and widely publicised to make things like changing your user agent incredibly easy? Eventually, there will be enough users doing it to really affect your site. In that case, there are a couple of options:
If I come across any other ideas, I’ll add to this post.
I’m pleased to report that after a trip to Andalucia, home of many of the finest classical guitar makers (and hence the lack of posting here lately), I returned to find that the SEO work on Kent Guitar Classics‘ site is beginning to pay off.
For some of the key target terms, such as “classical guitar dealers” and “vintage classical guitars”, huge improvements have been made, from obscurity and into the top 10 positions on Google.co.uk and Google.com, including a couple of number 1 positions. Below is the graph for “classical guitar dealers”, showing a rise from 32 to 10 on Google.com and straight into number 1on Google.co.uk & Live.com.
Just goes to show that a bit of keyword research and straightforward page optimisation can have a big impact!
An interesting post and comments at Search Engine Roundtable regarding Google’s recent statement on re-writing URLs.
Google has somewhat changed its mind about re-writing URLs, as they now claim to be better able to understand dynamic URLs (the sort of query strings you often see in e-commerce website addresses, for instance, along with many content management systems). The reason is that they now see query strings such as “search.php?keyword=toys” as more meaningful to the page’s intention and content than “search.php/keyword/toys”, which is how many URLs are re-written. The structure of the former is now properly identified by Google as a search term, whereas previously it may have had little meaning. Converseley, the latter now looks like a page three layers deep in the site, but doesn’t necessarily represent a search query, so Google is less likely to identify the true purpose of that page.
My take on this is that if you are re-writing URLs from something meaningless such as “page.php?id=76″ to something meaningful like “page.php/seo-urls-still-good”, that still helps both the search engines and users to understand the contents of the page and I would continue to use it. If you are re-writing search queries like the “toys” example above, maybe you could try a few without the re-writes – but remember that you could lose the PageRank of the originals, so be sure to 301 re-direct the old URLs to the new ones (and update your sitemap accordingly!)
I was shown this in a client’s Webmaster Tools earlier this week:
We had carried out a number of 301 redirects on some of their pages, as for reasons known only to the original developer, a lot of pages had been created as sub-domains, which was causing duplicate content and indexing issues with Google.
What I wasn’t aware of, was that there isn’t any code in the site to auto-update the sitemap.xml file provided to Google Webmaster Tools. I hadn’t seen the error above before – clearly, Google is unhappy if too many of the URLs in your sitemap don’t match what it sees on the site. A lot of those URLs of co urse no longer exist (e.g. the sub-domains), so we have updated the sitemap using GSiteCrawler – it’s a bit techie, but it certainly does the job and can be scheduled to make regular updates with automatic FTP of the new sitemap.xml file.
So, if you’re making changes to your site, remember to update your sitemap.xml files!
